Keras学习---RNN模型建立篇

本例子是“IMDB sentiment classification task”，用单层LSTM实现。

1. 输入数据预处理
输入文本数据统一规整到长度maxlen=80个单词，为什么呢？

是不是长度太长时训练容易发散掉，这样就限制了记忆的长度了。
x_train = sequence.pad_sequences(x_train, maxlen=maxlen)

有没有动态的呢？因为输入的句子长度本身是动态长度的。

2. Embedding layer

代码中，max_features=20000对应的是词汇量大小。

关于Embedding Vector,如果是中文的，怎么处理呢？

在RNN训练过程中Embedding Vector是否也参与了训练呢？如何选择？

3. 关于RNN的模型架构的理解

Embedding Vector是128维的，隐层是128个节点（也可以是其他数值）。

Embedding Vector与隐层的节点是全连接的，隐层每个节点自带存储单元的。

在这个128节点的隐层之上，有一个Dense节点。这个Dense节点是和128节点的隐层全连接的。

Foward过程就是每次输入80个单词中的一个，直到最后一个单词输入结束，Dense节点最终的输出就是估计的Y值了

完整代码如下：

'''Trains a LSTM on the IMDB sentiment classification task.The dataset is actually too small for LSTM to be of any advantagecompared to simpler, much faster methods such as TF-IDF + LogReg.Notes:- RNNs are tricky. Choice of batch size is important,choice of loss and optimizer is critical, etc.Some configurations won't converge.- LSTM loss decrease patterns during training can be quite differentfrom what you see with CNNs/MLPs/etc.'''from __future__ import print_functionfrom keras.preprocessing import sequencefrom keras.models import Sequentialfrom keras.layers import Dense, Embeddingfrom keras.layers import LSTMfrom keras.datasets import imdbmax_features = 20000maxlen = 80  # cut texts after this number of words (among top max_features most common words)batch_size = 32print('Loading data...')(x_train, y_train), (x_test, y_test) = imdb.load_data(nb_words=max_features)print(len(x_train), 'train sequences')print(len(x_test), 'test sequences')print('Pad sequences (samples x time)')x_train = sequence.pad_sequences(x_train, maxlen=maxlen)x_test = sequence.pad_sequences(x_test, maxlen=maxlen)print('x_train shape:', x_train.shape)print('x_test shape:', x_test.shape)print('Build model...')model = Sequential()model.add(Embedding(max_features, 128))#model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2))model.add(LSTM(128))model.add(Dense(1, activation='sigmoid'))# try using different optimizers and different optimizer configsmodel.compile(loss='binary_crossentropy',              optimizer='adam',              metrics=['accuracy'])print('Train...')model.fit(x_train, y_train,          batch_size=batch_size,          nb_epoch=15,          validation_data=(x_test, y_test))score, acc = model.evaluate(x_test, y_test,                            batch_size=batch_size)print('Test score:', score)print('Test accuracy:', acc)

通过summary()得到的函数统计如下：

____________________________________________________________________________________________________Layer (type)                     Output Shape          Param #     Connected to====================================================================================================embedding_3 (Embedding)          (None, None, 128)     2560000     embedding_input_1[0][0]____________________________________________________________________________________________________lstm_1 (LSTM)                    (None, 128)           131584      embedding_3[0][0]____________________________________________________________________________________________________dense_9 (Dense)                  (None, 1)             129         lstm_1[0][0]====================================================================================================Total params: 2,691,713Trainable params: 2,691,713Non-trainable params: 0